Fintan, a language learning fan, wrote to me the following:
Hi David, I am a language learner and am impressed with your Goldlist system. I want to ask you a question about the Goldlist method if that’s ok. Why did you choose 2 weeks as the minimum period for the first distillation? Was it efficiency/management of growing bronze lists that decided this period of time as I can’t find info on the association between 2 weeks and the end of short term memory? I understand the 20-30% retention by this passed time but I just want to check with you. Also, it occurred to me that one could activate this first 20-30% after only 2 weeks rather than waiting to be finished all distillations. By activating, I mean using a full circle method like Luca Lampariello to translate the known passive sentences back into the target language. I think this does not contradict the long term memory goal as you are only activating known words. Looking forward to hearing your thoughts. Fintan.
My reply, as ever when someone asks a great question which others could really benefit from;
Can I write the answer as an article on huliganov.tv?
To which Fintan kindly responds:
Absolutely :) Looking forward to it! Thanks
So here goes…
Now and again in my writings (which anybody could be forgiven for not having read all of), I have spoken a few times about what the significance of the two week fermentation period in Goldlisting is.
However, I don’t think I’ve really said enough about it and so this question is a very useful one for those who would also like to know more about the topic and many readers no doubt will be interested in delving into this particular issue more closely just as Fintan is.
As you know, the Goldlist method is based upon the findings of Ebbinghaus, known as the father of the study of memory as well as the father of clinical psychology. Very little prior to Ebbinghaus’ work had been done on the human mind and memory using the scientific method, namely experimentation and the testing of various theories by performing logical experiments giving tangible evidence which can be repeated by other people. Ebbinghaus was groundbreaking because he did this for the first time in a very imaginative way – and in a way very interesting for polyglots. That is to say he taught himself nonsense words and measured how quickly he would forget them. One weakness about his method was that he used himself as the main subject, being too kind to inflict the rigours of learning nonsense for the pure sake of forgetting them while measuring the rate of doing so on his students.
If you want to know more about this pioneering scientist, a good place to start would be the Wikipedia articles on him and on his theories.
Reading them however you will probably notice that two weeks per se is not mentioned. And neither is the idea of there being a short-term memory which is for conscious and long-term memory which is for unconscious learning. You can look for these concepts in the work of Ebbinghaus in the way that you could look for the term Trinity in the Bible, or indeed the name of God in the book of Esther. These are things which are intrinsic, and go without saying to a degree. It is not possible to understand all the observations of fact without such a theory, and therefore it emerges from the work even if not explicitlt mentioned.
Of course, it is also possible that I’ve read into what Ebbinghaus did more than he ever intended, and if that is the case then of course on the one hand that would put me in a position of being less “scientific” than such methods as Supermemo by Wozniak or Anki which certainly do map onto the forgetting curve in a way which the Goldlist method does not, but for reasons I’m going to come to an moment that actually doesn’t matter.
Let me just grab a picture of the forgetting curve if I can. Bear with me one moment.
Right, here we go. Thanks to http://en.wikipedia.org/wiki/User:Icez for this, which he placed into the public domain.
As you will see, it is not that you remember everything well for two weeks and then get a sudden deterioration. Not according to this, anyway, and not according to the raw data Ebbinghaus made from his experiments with himself.
There are reasons that I will go into a little bit later on in this article that make me think that such phenomenon does exist. However I don’t know of any actual scientific evidence proving the observation is measurably true rather than simply anecdotal but I will give you all the reasons why I think that the two-week cut-off is more than just an arbitrary point in the forgetting curve. That’s coming up below in the article. Read the rest of this entry
One reader with the pleasant name of Marlon wrote in one comment recently the following great question, and thus coaxed me to impart some advanced goldlisting knowledge which I was keeping back for the book:
I am eager to start the Goldlist method. However, I need further clarification about scheduling. I read your post responding to Abdul some years back but I am still not sure how I can avoid distillations and new headlist overlapping. I do understand I could simply insert a batch of words and not use the step system, but it is not my desire to take that route.
I would prefer to use the step Goldlist method. I think I am most confused by time allotment. I decided to use the 20mins/25words/10min break format. When distilling to the first set (from 25 to 17), I believe you suggest use the same format, that is to use 20mins/25words/10mins. What about D2? Do I still need to use 20 mins to go from D1–>D2 (and D2—>D3)? In other words, do I perform as many distillations as possible after D1 is completed in the 20 min allotment? For example, Would it be prudent to distill maybe 2 sets from 17–>12 in one 20 min block?
I am looking forward to your response.
I will write the answer to this as a main article, partly because it’s a better way to get more readers to read it, and it is a good and useful topic for those who are using the Goldlist, and partly because I can use tables better in a new article than in a response.
I think it’s an excellent question, which shows that you’ve understood most of what I need you to understand in order to work successfully with the method.
I have in the past left people to fill in the blanks for this one themselves, as there are a number of ways in which you could fill in the blanks and they would all be good as long as the basic tenets are agreed to, and also I was leaving something back for the book, but just to give you an example of what works for me, imagine that you decide to do a project in which you have a good idea how many lines will be in the headlist in total, and lets say it’s going to be 3000 lines of headlist.
I would split that task into Batches, and each batch I give letters of the Alphabet, so Batch A, Batch B, etc.
Now because we want to avoid running into within two weeks of ourselves, as well as not have too long periods of not getting to review the same material (more than a quarter of a year is not necessarily harmful, but means you have little momentum, in practice, which can be demotivating) we need to plan it so that the first batch is the biggest batch, and then they get gradually smaller.
So the last batch will be 100 words, the second from last will be 200 words, etc.
Now follow me through this logic: Read the rest of this entry
Reader Jarad Mayers wrote the following very good question:
I want to learn Mandarin. I am not sure how to go about it. This is the very first language I am attempting to learn. I have not done anything yet. I am on very tight budget and currently not employed. I tried to access the free material on Mandarin (http://fsi-language-courses.org/ )but it is no longer accessible . I was wondering if I could use your experince and if possible sort of outline the steps I need to follow.
BTW, I am not sure where to post my question. I am sorry if this the wrong place for posting it.
I’ve prepared the answer as a table – it is a whole programme to 80% of Chinese that you’d need to get your degree, read newspapers, live an everyday life in China. The rest after that comes down to vocabulary building for which I’d recommend the goldlisting of dictionaries or of bilingual literature. You could spend four times as much time getting from 80% Chinese to 100% Chinese (ask Vilf “the Gilf” Pareto, he’ll tell you why, or might have done, until 1923 – now you’ll have to look up what he thought in order to know why, or simply accept it).
Real Chinese philologists like Victor Berrjod might give you other useful sources better than the ones I have listed. All of the ones I have listed are available on Amazon. The audio courses are expensive so it will pay you to shop around a bit.
Many thanks to those of you who have subscribed and who come again and again to my humble abode. This milestone is hopefully only the beginning, although in fact I have now been doing this since November 2009, that is three years, and a few days on top. In three years’ time where I’d really like to be is over a million views – that’ll need a lot of work to do pulling over from other places all the resources and creativity that I’ve been doing in different parts of the web and making this the unified place where it’s all easy to find in one place with the various categories and subcategories, and the ability to search by words within this space, as well as the ability to have discussions not hampered by word limits, in which you can thread them properly and include links and media to your hearts’ content – unlike in YT where most of my material currently is and where most of my hits currently occur – in total well over 4 million there so hopefully a million here by the end of a similar six years (the time I’ve been on YT is now closer seven than six) is not too much to hope for.
In the end it depends on you, the viewer. Every bit of interactivity that you do here, discussing with me or with other commentors if you feel the urge, every subscription, every use of the share buttons I’ve put under the articles, it all helps me along, it all encourages me to produce more in the future.
Not everyone will like the blog, or the films and other internet “assets” (sometimes “internet contingent liabilities” might be a better phrase), but for some of you I know it has been and will be a source of interesting ideas and an experience of language learning, travel and other subjects such as faith, politics and others from time to time, and I hope that it will continue to be a place that you subscribe to, that you like to come to from time to time, and that you recommend to like-minded people. Read the rest of this entry
Victor asked to see some of the goldlisting of the Heisig book recently described in practice. I picked a relatively early point in the book to show – this is the 10% mark – and please note in this particular book the headlist is on the right and the first distillation on the left. No frames were actually distilled out on this page but you can see the stories getting shorter.
Mugiwara wrote some very good questions which deserve a reply as a new article. I have also today answered smaller but equally good questions on the Goldlist Methodology page, so people with outstanding queries about the method may also like to read them. Anyway, here goes for Mugiwara’s great questions:
Hi there Mr. Huliganov.
I’m Spanish and I’m trying to learn Japanese, this language seems complicated using Gold List Method because of the kanji but I have some basic questions because my English skills are not good enough and I don’t understand some points of the method.
Kanji, especially when done the Japanese way where you have usully at least two readings as opposed to Mandarin where there is usually one and sometimes two and greater phonetic clues are embedded in the primitives for Mandarin than for Japanese, is not possible to study in exactly the same way using goldlist. The ideas behind Goldlist still hold true, but they need to be applied in a different way and the task of mastering Kanji, and Japanese in total, has to be broken down into a jigsaw, each piece of which needs to be mastered as a piece and then put back together again.
The people who gave us sudoku, sushi, bonsai trees, origami, manifold management techniques and martial arts aplenty have actually set us the most subtle and challenging puzzle of all in the form of their own language. As with all things Japanese it takes a certain technique to get it right. With the technique it is possible, without the technique it seems impossibly difficult and unachievable – still beautiful, but remote and not fully understood. That seems to be par for the course with everything they have.
In going through the answers to your excellent questions today, I will try to make clear how I think the ideas of Goldlist can best be reapplied to the question of kanji learning, which in itself is only part of Japanese language learning in total. Even when we know the Kanji and their readings, it is necessary to know the combinations and just as the Kanji themselves run beyond 3000 (of which less than 2000 are in the obligatory lists of the Ministry of Education) so the combinations of them run into the tens of thousands, and very often a word we want will not be a single kanji. We think of words as words but if you take a series of words, lets say some different metals and alloys as one series, or some well-known birds or reptiles as another, some plant types as another, let’s say, we’ll find that sometimes a word in English will be a single kanji word in Japanese, the next in the series may be a two-kanji combination or even more than two, and then the next may be a word not written in kanjis at all but may need to be written in katakana because it counts as a loanword or a technical word or an onomatopoeia. But we will not learn to run before we can walk.
1 – I read people is trying to do huge lists like 600, 1000, 2000.. and that sounds a little scary so, as a beginner in your method, how many words are recommended to familiarize yourself with the method?
I think it is good to do first off a batch of 500 words, but if we were talking about kanji I would be shaping the method rather differently. I suggest you might take the kanji which are usually listed for JPLT #5 and there are between 180 and 280 of these depending on whose book you read or whose website you visit. I do highly rate Heisig’s 3 tome oeuvre, which is also available in Spanish although you certainly don’t need it, and a first batch could be just the “part one” kanji from that book, which is not very many. The important thing is – and this is what Heisig says and many Heisig readers seem to ignore it – you need to study the kanji with a pencil or pen in your hand and draw the kanji while thinkning of the story, but you don’t need to write it over and over at the same time. This traditional approach to Chinese characters involves a lot of wastage of time.
So to apply the Goldlist method to kanjis using a source like Heisig, instead of having the usual line by line approach, I take a book and freely write out the following: The frame number per Heisig part 1, and the meaning (and you have to be very precise and not paraphrase Heisig’s meanings). Then I write the component primitives and an outline of the story that links the primitives to the kanji meaning. I do the same for primitives that have no Kanji status too, but they have no frame number, just an asterix.
I only use one side of the page and leave the other side for revision, and I might get five or six, or maybe only three or two on a page, depending on how much there is to say about them. At any rate, I probably wouldn’t write the kanji itself in full out more than three times. I might write the stroke order if it isn’t obvious.
I just try to go through the thought process James Heisig is presenting for the given kanji, or making one of my own up if I can see a clearer one for me than the JH one, and I write it down so that all the info is in the “Headlist” in my book and I don’t have to refer back to the book very much when I’m reviewing.
I go to the point where I have been doing it for two weeks at least, probably three or four, as otherwise whe I revise I will run up against the minimum time rule of the Goldlist method, and by that time I’ve probably done 250 or so, just by doing an hour or so every other evening. I will probably have filled the right or left hand pages of an 80-96 page A5 format writing book with these 250 kanji, in other words I dedicated to them more book space than 2500 words that I could just write out phonetically. You cannot easily learn kanji in my opinion on a one-line-per-kanji basis. It is possible to do some things with kanji that way but I would leave it until I had the understanding of kanji as kanji and then maybe do combinations, the various yomi and maybe practice sentences that way. For getting used to writing and making long term associations for primitives and how they fit into kanji and what the base meanings are, I need a much freer format, but I still have certain truths from the Goldlist method which can be brought into to service this situation.
Therefore after the requisite time of at least two weeks and a buffer on top so that I don’t catch up with less than two weeks of myself in the middle of a batch, I use the other side in the book to do a very similar thing to what I would do in the traditional Goldlist method, namely I’m going through the material on the one side, seeing if I now know it, leaving out the ones where I can write the kanji with proper stroke order and know the meaning of it both as kanji and primitive per Heisig’s method, and the way it will appear as a primitive when pressed to the henben or left side position or the crown position or the bottom position. I do not need to know the sound in order to drop them, as in the Heisig method readings come later. If I were learning Mandarin I would probably want to know the Pinyin and have learned that, and Hoenig’s book which I use in preference to Heisig for Mandarin does have them at the same time, but that is an awful lot to want your memory to do at once and maybe it isn’t the best idea to try to do that. I don’t want to talk about Mandarin here when you are asking about Japanese, but there may well be something to be said for taking a Heisig book one approach to the Chinese characters as far as characters are concerned, do the grammar and get used to the language itself using Michel Thomas primarily followed by Pimsleur (which are audio only) and then bring in the pinyin. That is certainly the best way forward when it comes to Japanese. Even the pinyin or roomaji writing, which can be helpful of course, don’t need to come in until after one has worked through a good 12 or 15 hours of structured audio learning of a high quality, like the MT Japanese course.
That was a bit of an aside, so back to your actual questions:
2 – If you are going to do a huge list, suposedly you have to write 25 and then take a short break like 15 minutes, ok, but then you need one week or more to write all the words, right?
Let’s imagine now that you had a list of words which were all katakana words which you wanted to put to your long-term memory – you could do it on the usual list of 25 way, as is usual for the Gold list. Or if you were learning Japanese in Romaji (which can be one way of breaking it down, but I don’t recommend making do with just Romaji, doing that would just be one part of the puzzle) then you could do a 25 word list in the usual way. You would choose a batch size like 500 as above, or whatever your word list was that you wanted to learn.
Let’s say someone gives you a list of katakana words, let’s say the top few hundred by frequency words properly written in Katakana in Japanese, you could really, as long as you were comfortable enough in katakana, go ahead and use katakana to learn them in the usual goldlist manner. You would take maybe 20 – 25 minutes (depending on how well prepared the source was, and how fast you work when working for maximum comfort and enjoyment) to write your head list per double page of 25 words, and then you would probably go away for 15 minutes and do something completely different just to rest that unconscious function. If you don’t then you could have it giving way to consious learning attempts and short-term memory functions without you even realising what was going on. After all, the thing about the unconscious is that it’s not conscious, so we don’t feel it. We know it’s there because it’s also what keeps us breathing and our hearts beating, etc, but usually we ignore it and that’s when it does its job best.
3 – After you create your headlist and let’s say a month later, do you just try to write out the words your remember or you look your list in your language and translate it?
The list has both languages in it, usually (unless the meaning is obvious and I’m just remembering the spelling or some perculiarity about the word other than its meaning), my target language and the language used for learning, which will either be my own language or a language I know much better than the target and I’m learning via that language either to hone it or because the second language is in the same group and so I’m using materials made for speakers of the first language in the group that I know, as this will home in on the differences between those two languages and reduce my risk of confusion and linguistic interference.
Now when I am reviewing it again all I want to do is objectively ask myself do I remember it or not. It is not a question of being able to go from your language to the foreign language – this is too high an expectation and anyone who expects that is chasing ends of linguistic rainbows. It is sufficient to ask yourself whether you remembered the meaning from the language you are learning into your own language or the transit language you are using to learn this new language. On top of the meaning you can ask yourself “would I have remembered the spelling” or “would I have remembered the grammatical irregularities” or “would I be able to pronounce it” whatever the reason was (when it comes to later distillations especially) that you didn’t take it out of the list earlier.
Beyond flatly taking out, you can also very validly combine some items into single line items.
Either way, if you know a word, you won’t write it again.
Now let’s go back to the idea of kanji. If we are following Heisig’s books and goldlisting them, we will consider a primitive or a kanji learned for the purposes of book one when we know the meaning of it as Kanji and Primitive, when we know the stroke order and variations of it as primitives in different positions. We need to be sure we can tell the difference between this primitive/kanji and similar ones. If we are sure that we have that image of the little story so that we really recognise the kanji or primitive and can give its meaning as in the book, would recognise it as part of another kanji with fresh elements and can write it out confident about stroke order then for the purpose of that goldlist it is learned and you can drop it.
It doesn’t matter that we haven’t got on to readings yet. Heisig students do book one to the end and then they do book two.
And it is by breaking it down this way that it becomes possble.
My Kanji goldlist bronze book has only two sides instead of the four sides I use for word goldlisting. Less detail from the stories need to be repeated in the later distillations so when it comes to the second distillation and I have a new book the silver book, I’m able to put 5-6 per side on average and as these aren’t more than about half of the ones I set out with (I do put every single one into the headlist) I need to put in a consequent number in addition to the Heisig book one frame number. I am not finished doing even the Headlist but it is going well so far, and I know that I will need 8 bronze books of 80-96 sides A5 each for H and D1, and 4 similarly sized books for D2-D3 and 2 of these books for D4-D5 and just the one for D6-D7.
I work using a sort of batch-step method where I take that first batch through to the end of H, then I go back and take it to D1 and afterwards add batch#2 at headlist at the end of batch 1 headlist.
After that I take batch #1 from D1 to D2, take batch #2 from H to D1, and add batch #3 after the end of it.
Let me show you that pictorially:
Here is a plan by batches and distillations of how to get through the Heising book one. A person could put on the planned time or date and also afterwards show the actual and the actual work revised if they wanted to, for each chunk.
Now let’s use the order of colours in the rainbow to show the order in which I’d take each part of the work, that is each cell in this plan.
I’ve used pixel heights for the rows of work here that exactly correspond to the % still included in the work, so that graphically this shows very clearly what the work is in total for a good approach to a big Goldlist project. To learn 2040 kanji you do 7350 pieces of work, that’s an average of three and a half per kanji – actually in line with the results of Ebbinghaus, Wozniak and most other long term memory exponents, which is no surprise – Goldlist works to your biology, it doesn’t change it. Planless repetition would give you actually a much higher workload, and many people who embark on such an exercise never come to the end of it.
You can see that each sweep of the grid using a plan like this gets progressively longer until the end of the material is reached or the end of the planned number of distillations is reached, which in this 8*8 arrangement happens at the same time. I’ll call each sweep or cycle a “pass”. In the first pass we only have the red cell so that is 240 items. Then we take the pass of the orange and yellow cells and that pass takes 448 items, so you need to make more time for it. The next pass where you have green, blue and violet is already 598 items and the one after that has a nice round 700 items, and so it goes on until the biggest one, the eighth pass, which has in this case 917 items, and then they quickly fall, so that the 9th, 10th, 11th and 12th passes have 664, 471, 327 and 220 respectively as you can see from the table if you add them up, and after that point you start to have the problem that there isn’t enough to do in the two weeks you are supposed to leave between reviewing the same work, although in practice at that point it’s safe to be concurrently working on the Second Heisig Book with a separate project anyway, which you would run on similar lines.
If each item takes 2 minutes on average, which they should if you bear in mind that we write less per item in the later distillations, then this whole project is a question of 245 hours of study, while you’d break up into chunks with the breaks so that you would certainly need to take a while over it.
Allowing that there are 15 passes but that you can do concurrently the next phase after 12 of them, I’m saying that the minimum time that I’d recommend giving learning Heisig’s first book is 24 weeks, that gives you your “mandatory” two weeks per pass in order not to bump yourself. The middle parts of that need though for you to be doing according to this logic some 500 items a week which is 1000 minutes or 18 hours work a week, but the average workload of 245 hours over 24 weeks is clearly 10 hours, which is a good deal lower. What’s more likely to happen is that you’ll have a bit slower progress in this big passes. And then you need to give a similar length of study to Book two in order to get all the kun and on yomies learned, as a separate issue.
Hence learning the joyo kanjies, their meaning and their readings before you even start to use them in sentences is a year’s work minimum. If you can do if faster your own way, then fine, but I can tell you that it means in most cases a good deal more than the 500 hours more or less I’ve suggested here to work through Heisig one and two with Goldlist principles.
4 – In your explanation of the steps to Taylor, you did a “new step” which is like creating a new list in the middle of the other one, with new words I guess but when did you started it?how long after the second destilation? and then you do two destilations at the same time? I’m a little confused.
This step is not obligatory, but can be useful if you are going slowly because you are busy with other things. I will talk more about it in the book. Don’t worry about that step for now.
I hope you can understand my questions because my English skills are just decent, and thanks.
Your English skills are more than just decent they are superior, at least from the writing I’ve seen, to the bulk of native speakers. If you achieve the same in Japanese that really will be impressive, and you might, if you work with patience, stamina and a good method! Many thanks for the great questions which I believe will have helped others also.